60 Hz waveform distortion

Status
Not open for further replies.
160111-0944 EST

Meternerd:

I share a 50 kVA transformer with one neighbor and two street lights as the load. This transformer was new several years ago because the original 45 year old transformer was leaking oil.

When the voltage waveforms were taken the neighbors were out of town. I believe our substation is a wye source. But our loading is effectively delta, and several homes have open delta loading. Our transformers are bridged line to line and there is no neutral on the poles.

I believe this is a system wide voltage distortion. At some time I will look at voltage on a different substation.


SG-1:

Was your waveform taken at home or at your lab? Does your scope have the ability to average the signal, thus reducing noise and improving the apparent resolution. I have used various average counts in the above images. Not less than 2 and I don't believe more than 64.

.
 
The waveform was from home, taken at the power strip that supplies two laptops. The scope is a very basic scopemeter < $400. It was taking 10 samples per second, adjustable in the software.
 
Here is my contribution

Home, taken at SE
Puget Sound Energy
5 homes ranging from 5300 sq ft to 2800 sq ft on the 35 kVA pad mounted transformer fed with 5.1 kV underground service. (PSE contractor, Potelco, believes in using xfmrs at or near their 150% limit). No idea what the other homes are drawing, but believe 3 of the other homes are on natural gas vs. heat pump, and no known heavy loads. All homes have 200 A 120/240V service.

At time of reading, in own home a 2T heat pump operating and 3 computers, possible a 1.5 HP water pump, minimal lighting load.
TEK 2430 storage scope, 10X probe, single sweep (not averaged). Obvious 3rd harmonics and a few volts of zero sequence offset. Looks like a few mV of 20 kHz electronic ballast noise. 90 feet of underground 4/0 Al feeder plus 80 feet of 2/0 Al aerial feeder to SE.

DSCN7406.jpgDSCN7407.jpg

edit, no idea how the bottom image came to be - does not show up on edit page to delete.
 

Attachments

  • DSCN7405.jpg
    DSCN7405.jpg
    133.5 KB · Views: 3
Last edited:
160111-1536 EST

Junkhound:

Thanks. Your waveform has a very similar peak distortion to my waveform. But it does not show the inner curve to the right side of the peak that my plot shows. I assume that Boeing and Microsoft are a load on your system.

I don't believe the distortion is a result of a local distribution transformer, but rather is a composite result of all the users on a system.

My guess as to the cause of the waveform you show is a very large number of capacitor input filter power supplies. I can lower the total load on my pole transformer to virtually no load and the same waveform distortion exists.

.
 
More like capacitors on the output of thousands of single phase rectifiers are the culprit.

One of the office buildings I worked in during the early 1990s was supplied by a large 3 phase delta-wye transformer on the roof, installed in 1986.

The company 'upgraded' everyone's computer in the 90's (forget the computer name, an early Dell) during the late spring.
First hot day (hot here is over 80F) in summer the transformer primary winding fried.

The power supplies in the computers all had single phase input full wave bridge rectifiers with a 500 uF or so capacitor, no PFC circuits yet in those days. Triplett harmonics from all the new computers were enough that the circulating triplets in the delta primary were enough to release the 'magic smoke'. Had to size the replacement xfmr for the triplett primary heating rather than for load.
 
gar said:
160104-2137 EST


Although I have three phase on the pole I don't have it in the house. Thus, I don't know how the other two phases look here.

If you have three phase on the pole, is your service 240/120 single phase or 120V/208 network? Either way, the POCO could move you to different phases to see if the problem improves. If it doesn't, then there's a three phase service somewhere that may be putting those harmonics on the line. POCO's don't like loads that put "noise" back on their lines, so maybe they'd be willing to work with you to try and find the source of the problem.
Click to expand...
 
junkhound said:
Here is my contribution

Home, taken at SE
Puget Sound Energy
5 homes ranging from 5300 sq ft to 2800 sq ft on the 35 kVA pad mounted transformer fed with 5.1 kV underground service. (PSE contractor, Potelco, believes in using xfmrs at or near their 150% limit). No idea what the other homes are drawing, but believe 3 of the other homes are on natural gas vs. heat pump, and no known heavy loads. All homes have 200 A 120/240V service.

At time of reading, in own home a 2T heat pump operating and 3 computers, possible a 1.5 HP water pump, minimal lighting load.
TEK 2430 storage scope, 10X probe, single sweep (not averaged). Obvious 3rd harmonics and a few volts of zero sequence offset. Looks like a few mV of 20 kHz electronic ballast noise. 90 feet of underground 4/0 Al feeder plus 80 feet of 2/0 Al aerial feeder to SE.
Click to expand...
The asymmetry suggests even harmonic numbers.
 
Besoeker said:
Not sure. I think that would affect one half cycle rather than one half of each half cycle.
Click to expand...
Right. That would actually be a subharmonic rather than harmonic distortion. I was not thinking clearly.
Possibly SCR with the right trigger or cutoff angles? More than linear load current near the zero crossings rather than near the voltage peaks?
Combined with reactive loads?
Not at all sure what could cause that behavior in the real world.
 
junkhound:

That bottom image was done using the Attachments process. It looks like a paper clip.

Your other 2 images were done using the Insert Image process.
 
GoldDigger said:
Right. That would actually be a subharmonic rather than harmonic distortion. I was not thinking clearly.
Possibly SCR with the right trigger or cutoff angles? More than linear load current near the zero crossings rather than near the voltage peaks?
Combined with reactive loads?
Not at all sure what could cause that behavior in the real world.
Click to expand...
A leading edge dimmer could.
But it is not subharmonic. It repeats with each cycle.
 
Not at all sure what could cause that behavior in the real world

I did a really fast check just to put up a waveform.
I notice that the scope is still set on DC coupling from some previous work.
Thus the offset may be due to instrumentation error, although the particular scope used is on an isolation transformer with the scope chassis ungrounded. Note that the zero is not right on zero so the apparent offset of the 60Hz is not as great as it appears, but still is some.
Who knows, might check sunspot activity today - back in 'whenever' (1990s?) Ontario hydro lost a couple of 3 MVA transformers due to an MHD 'dc' (a few millihertz) offset causing saturation at peaks.
 
sloppy tech work

sloppy tech work

my apologies, no offset - just me being sloppy in a quick setup. Spaceweather.com says minimal sunspots.

156.0 volts peak both positive and negative.

Same conditions as before, just different scope setup (and crooked camera), a few hours later, few more neighborhood loads on so voltage lower.

dist2.jpg
 

Attachments

  • DSCN7408.jpg
    DSCN7408.jpg
    130.9 KB · Views: 1
  • dist.jpg
    dist.jpg
    19.7 KB · Views: 1
160111-2156 EST

Some newplots.

This morning at midnight, 0000 hours, at home I obtained the following plot with virtually no non-linear loads on in the house.

DS2_QuickPrint32-A.jpg


The blue curve is the full AC line voltage. There appears to be less distortion than in previous plots. i did not capture an expand plot of the peak. At this time of night on a Sunday night I would expect much of Detroit industry has not yet started up.

The red curve is the neutral to EGC voltage at my bench without filtering resulting from the RIGOL input current. The curve is flat to less than 10 millivolts when the AC power to the scope is derived from a different source, an extension cord to a different outlet that is not on the bench circuit.

This post was to check out a technique to get a better plot picture. Next post has plots from the shop with three CNC machines running. Quite interesting.

.
 
160111-2024 EST

More plots.

This morning at 0930 the following measurements were made. Detroit industry is fully operational, as are our local Universities and Hospitals.

The first plot is the full AC voltage and note the increased distortion as compared to last night at midnight.

View attachment 14140



The second plot is the same time, but vertically expaned near the peak.

View attachment 14142

Note: the considerable distortion with miscellaneous humps and bumps.


The next plots are late in the afternoon near 5 PM at the shop. Three or four CNC machines were running. The power source is an open delta with one transformer center tapped for single phase loads. The voltages displayed below are from 1/2 of the center tapped secondary. I thought I associated the big change in distortion as the CNC lathe cycled thru its part. Lathes usually have a much higher inertia load on the spindle than do mills. Even on parts that require little load power there is a long power load to run the spindle up to speed.

In this case the local load, CNCs, are the primary contributer to the distortion you see. I would have to look at this during a down time to see what distortion is from the power company primary. This location is 2 miles from my home, and on a different substation. As the CNC machines cycled thru their operations it was interesting to watch the variation in distortion.

There is only 1 or 2 minutes time difference between the following plots.

View attachment 14142


Next:

View attachment 14143


Next:

View attachment 14145




Next:
DS2_QuickPrint40-A.jpg


Next:
 
160111-2509 EST

One can create something that works correctly in preview mode but is screwed up when saved to the form. The above can be viewed by clicking on attachment. But it is a mess to go from one plot to the next and comparisons are difficult.

.
 
gar said:
160111-2509 EST

One can create something that works correctly in preview mode but is screwed up when saved to the form. The above can be viewed by clicking on attachment. But it is a mess to go from one plot to the next and comparisons are difficult.

.
Click to expand...

I'm getting an "invalid attachment" message when I click on your links.
 
160112-1127 EST

gadfly56:

This forum software has bugs. I have no control of that and could not make any corrections. Nor do I have any idea of how to correct the problems. Everything worked correctly in the preview mode before I submitted the post. After posting I could not make corrections. All of the plots should have directly displayed, only some did. All of the plots should have been stored within the forum memory. All of those that required clicking on should have displayed, but only one worked.

Having to click on a link makes it hard to compare plots.

I may have to go back to putting pictures and plots on my website where I have more control, but it is a lot of extra work.

.
 
Status
Not open for further replies.
Top