Erratic Voltage Distortion At Zero-Crossing Point??

Status
Not open for further replies.
B

Baeljb

Member
Location
Denver, CO, USA
​I am troubleshooting a flickering light problem at a one-year old residence and would appreciate suggestions from members. The residence has 70+ assorted electronic dimmers serving a variety of lights including LED, incandescent, etc. It has a 1000amp. 120/240 single-phase service. At lower output (50%) the lights will strobe multiple times per second, for hours at a time. We, and the dimmer manufacturer, have done extensive testing with high-end meters and storage oscilloscopes. We have concluded that there is erratic (shifting from cycle-to-cycle) voltage distortion on the utility feed. RMS voltage is jumping 0.2 volts per cycle. It is present when all loads in the home are off and also present at neighbor’s homes and a local business two miles away. The distortion is present at the zero-crossing point. The dimmer manufacturer says that they reference the zero-crossing point to adjust output voltage. They do not have a fix. I would normally expect to see Vthd on the service but not have it shifting once large non-linear loads on the distribution stabilize. Vthd is less than 4% and within the utility standards. We are attempting to pressure the reluctant utility into investigate their feeder loads and would appreciate any feedback on possible causes. I should mention that once the residence is transferred to its emergency generator, the flickering disappears. Thank you!
 
180726-2431 EDT

Baeljb:

Can you post pictures of the waveforms?

As an experiment try a dimmer that requires 3 wires, neutral, hot in, and hot out. I call them 3 wire dimmers, but may actually have more wires. These dimmers use the input hot and neutral to power the dimmer electronics. This is in contrast to a 2 wire dimmer where power to the dimmer electronics is derived from the series circuit thru the dimmer.

The nature of the zero crossing distortion will determine how likely it is the cause of your pulsation or flicker.

A 0.2 V RMS variation in supply voltage is very small. See plots P26, and P27 at http://beta-a2.com/EE-photos.html . P27 is just after midnight and a shorted time frame. P26 is a full 24 hours.

.
 
The dimmers all have neutrals, so hot and neutral in. Hot out and then neutral out spliced to neutral in. The dimmer manufacturer sent two engineers from the factory, took their readings, went back to their lab and recreated the waveforms and concluded it was the shift at the zero-crossing point not the 0.2 volts. They also brought out a number of different dimmers to unsuccessfully try. I have a nice video .wmv file showing the shift from one cycle to the next, but can't figure out how to post (ideas). Attached waveform pdf is from the manufacturer showing the shift from their scope. The second is from our meter showing about 4-hours of frequency. The meter uses the zero crossing point to calculate frequency. You can see a portion, when the home is on generator that the frequency stabilizes except for load steps.
 
Last edited:
Having some problems with attachments. Here are three.
 

Attachments

  • Dimmer Freq.pdf
    78.4 KB · Views: 1
  • Waveforms 5-31-2018 2.0.pdf
    121.3 KB · Views: 2
  • Waveforms 5-31-2018 2.1.pdf
    97.8 KB · Views: 1
My first guess would be that it is not spiky enough to be a bad connection. It also is steady enough to probably not be single-event switching.

Looks like an up-line harmonic. Some other load on the system is introducing noise.

I would see about getting the utility to investigate as they will have to get other loads dropped to find the source.
 
180727-1359 EDT

Baeljb:

Plots are are good information.

When I look at my voltage I have never seen anything like the bad quality shown in your plots. My voltage has distortion near the peak, but my curve is smooth. 60 years ago I had a near perfect sine wave. My present distortion results from the millions of personal computer and similar loads on the system.

In your plots I see a very high system load that turns on at about 60 degrees and off about 120 degrees. This hugh load pulse could impulse excite various resonances in the distribution that cause all the abnormalities in the voltage waveform.

You certainly have a power company problem. A low pass filter on each phase coming into the home might reduce the problem. I have no idea of the cost, or effectiveness.

My voltage at one time, see photo P5 at
http://beta-a2.com/EE-photos.html .

.
 
180727-1607 EDT

Looking more closely at the curves.

The first curve seems to have turn off at about 105 degrees on the positive peak. The negative peak is similar to the positive, but turn off is not as obvious.

It almost looks like a large load is switched on and off around the negative slope zero crossing, but not symmetrical about 0 V. Does not appear on the positive slope zero crossing. I can not tell about the positive zero crossing, does not look like it is there..

The second curve looks like the first except for the turn off point. The second curve shows positive and negative peaks being similar in turn on and off points.

There does not appear to be a shifting in positive zero crossing positions.

.



.
 
180727-1716 EDT

Connected up my 16 bit resolution digital scope to look at my present line voltage..

I see nothing of the garbage on my waveform that is shown in the previous plots. I see my peak waveform distortion, and a small perturbation, a short tiny 1 cycle transient at a high frequency, at turn on of a dimmer cycle.

The FFT shows fundamental at +10 db, 3rd at -40, 5 th at -30, 7th at -35, 9 th at -40. and everything else lower. Minor peaks about 1020 Hz, and 1140 Hz at -50 db. Noise level is just below -50 db.

To have the problem shown in the previous plots I would guess there is a moderate common impedance to both the cause and the residence. The cause of the distortion is some facility using phase shift control of current at high currents. Possibly electric arc furnace, chemical processing, aluminum ore processing, etc.

What kind of heavy industry is nearby?

..
 
Sorry for slow response

Sorry for slow response

I was out fly fishing on the Colorado River this afternoon. Have to set priorities. To answer some questions. I am attaching a spreadsheet with some interesting data. It can log every 8msec./ 1/2-cycle. We are looking at input/output of a 120vac, 1000watt dimmer. The voltage fluctuation is actually amplified by the dimmer. The residence is at a Colorado Ski Area. I know! Distortion still present once ski lifts and snow making equipment turned off at end of season. The resort is about 10 miles away from a molybdenum mine. The mine is a primary utility customer with their own substation. There is a water treatment plant for the resort and a water treatment plant for the mine tailings (don't know if it is on mine sub or residence feeder. No other big loads. I would not be surprised if VFDs in the plants were creating distortion, but I would expect it to be consistent from cycle-cycle once motors are up and running. It has been suggested, by a fellow PQ consultant, that there may be a resonance on the feeder, or a "bad" VFD. Also, maybe a bad capacitor on the front tend of VFD. Trying to figure out how to clarify. Thanks!
 

Attachments

  • Test Summary 4-05-2018.pdf
    61 KB · Views: 0
180727-2124 EDT

Your input voltage is extremely stable, although a very short time.

Do you have a special instrument that can measure an RMS value quantized over 1/2 cycle periods. In other words for each half cycle, synced to the zero crossings, is the instantaneous voltage measured, squared, averaged over the half cycle, and then square rooted to get a single RMS value for that half cycle?

Frequency looks very stable. But how was frequency measured? Was it on a half cycle basis? In other words the inverse of measured time for 1/2 cycle.

Output RMS voltage is much less stable than the input RMS voltage. I go back to my earlier question of whether the dimmer is really a 3 terminal dimmer.

The big industrial plants might well be the cause of the problem.

VFDs probably cause a large peak load current in the 90 to 150 degree range, and the angular range will vary as load changes.

Unlikely you can get power company to make any changes. I suspect transmission line lengths, loading coils if any, and capacitors are causing resonances that are excited by industrial plant loadings.

Again I question how the dimmers work. Are they really a 3 wire dimmer?

.
 
Dimmer Module Wiring

Dimmer Module Wiring

I am sure that they are using 3-wire modules and dimmers. That was one of the first items that I verified. I am attaching the installation information for the modules. Standalone dimmer instructions are the same.
 

Attachments

  • Dimmer Module Wiring 1.pdf
    133.3 KB · Views: 0
Baeljb said:
​I am troubleshooting a flickering light problem at a one-year old residence and would appreciate suggestions from members. The residence has 70+ assorted electronic dimmers serving a variety of lights including LED, incandescent, etc. It has a 1000amp. 120/240 single-phase service. At lower output (50%) the lights will strobe multiple times per second, for hours at a time. We, and the dimmer manufacturer, have done extensive testing with high-end meters and storage oscilloscopes. We have concluded that there is erratic (shifting from cycle-to-cycle) voltage distortion on the utility feed. RMS voltage is jumping 0.2 volts per cycle. It is present when all loads in the home are off and also present at neighbor’s homes and a local business two miles away. The distortion is present at the zero-crossing point. The dimmer manufacturer says that they reference the zero-crossing point to adjust output voltage. They do not have a fix. I would normally expect to see Vthd on the service but not have it shifting once large non-linear loads on the distribution stabilize. Vthd is less than 4% and within the utility standards. We are attempting to pressure the reluctant utility into investigate their feeder loads and would appreciate any feedback on possible causes. I should mention that once the residence is transferred to its emergency generator, the flickering disappears. Thank you!
Click to expand...

Not directly relevant but we had problems with variable speed DC drives in a paper mill. These used the zero crossing as a reference for the firing angle. The manufacturers of the drives were of no help. We fixed it by puting a small filter inn the sine wave signal that was being used to generate the firing angle. It was a long time ago but I think the filter might have been 0.5ms.
 
Status
Not open for further replies.
Top