Show us your mains waveform!

[Simon]

I'm not attacking you. I am stating for the umpteenth time that you are making a wrong assumption that you keep repeating. So lets do it the other way around. Why do you think turning a few hundred watts off in your home compared to the MW of power been used around you is going to make a difference? You seem so sure so please explain.

[engrguy42]

I'm not attacking you. I am stating for the umpteenth time that you are making a wrong assumption that you keep repeating. So lets do it the other way around. Why do you think turning a few hundred watts off in your home compared to the MW of power been used around you is going to make a difference? You seem so sure so please explain.

Simon,
Have your lights ever dipped when you turn on a big load in your house? Does it make the lights dip in the supermarket in town?

[Simon]

We are not talking about voltage, we are talking about waveform distortion that comes in from the mains. By big load you are talking several kW and no I don't have any load like that, that is the point! A load that large would be restive or otherwise PFCed. But when your lights dip is it due to the tiny wiring in your house or does "your big load" cause a dip all the way back to the substation. Find me a 3kW non PFC load, i didn't think your 3 PC's used that much. You have already got the result but you refuse to believe it. You turned your loads off and no change, that is because the power already comes in with that distortion, it's not wholely caused by you.

In your analogy it's like saying the light dipped because I turned on a 3kW heater, so I'm going to reduce it to 2.95kW and the lights should come back up....

[joeqsmith]

Curious how the mains wave will look like when using hair dryer like this condition below where it generated a DC offset at the mains line. Did this measurements many years ago as I didn't have HV diff.probe yet at that time.

It was Philips 1600 watts hair dryer, more details

Wouldn't you think that they rectify the signal to reduce the power and run it without for the highest?  Wouldn't you think that the lines will have some resistance?  Wouldn't you expect to see a small bias? 

Imagine if you could bias the line, all the way back to the transformer....

[joeqsmith]

......none of the screenshots posted thus far are as bad as some I've seen.......

Tautech, I was hoping to see some of these waveforms you mention.   If you have not captured them as I have done,

https://www.eevblog.com/forum/projects/show-us-your-mains-waveform!/msg3048418/#msg3048418

could you hand sketch what they look like?   When you start looking at power throughout the world, it's amazing how poor some of it is. 

[engrguy42]

Simon,
Perhaps the attached might help? A discussion of voltage distortion.

It gives a nice perspective of your house load relative to the "grid", and clearly states my point. 

[Jay_Diddy_B]

Hi,

Using the magic of LTspice I am going to show you where these waveforms come from.

In power systems engineering they use a system of 'Per Unit'. The concept is that at any point on a distribution grid if you load a point with its rated load the efficiency and regulation will be the same.

If you have a 10kVA distribution transformer and you load it with 10kVA the regulation and efficiency will be the same as if you have a 1kVA distribution and you load it with 1kVA.

Knowing this and making some assumptions we can estimate the impedance of the mains.

If we assume that distribution system is 98% efficient and 4% regulation with a 0.9PF load and working at 1kVA, knowing we scale to other power levels, we get:



If we run this simulation, there is nothing remarkable.
All waveforms are sinusoidal.
the Current lags the voltage by 25.8 Degrees (PF=0.9)
The load consumes 1kVA
The load dissipates 900W
18W is dissipated in the source

Mixed Load

If I change the load so we have

90% Resistive Load (900w)

10% Rectifier Loads (non PFC smps or LED lighting) 100VA

We can model this as:



And the modeling results are:





The results are remarkably similar to the measured waveforms shown by forum members.

If I perform an FFT on this I get:



The THD is 1.8%

Note that the 3rd and the 5th are largest harmonics.


This shows how if 10% of the loading is a rectifier load it can cause 1.8% THD in the voltage waveform.

Additional Note


The traditional method of PF = COS (Angle)

Because the waveforms are not sinusoidal this definition cannot be used on the rectifier load.

PF = Real Power / Apparent Power 

or Watt / VA

can be used

In this case PF = 67/100 = 0.67


Very typical for a rectifier load.

Regards,
Jay_Diddy_B

[engrguy42]

Wow, Jay_Diddy_B, now THAT is some good technical discussion. Thanks much.

I haven't gone thru it in detail, but FWIW, you mentioned estimating the mains impedance: the typical impedance of a distribution line is around 0.6 ohms inductive per mile for a 12kV line (at 60Hz).

[engrguy42]

Oh, and I recall a typical distribution transformer impedance is somewhere around 5% (ie, 5% voltage drop at rated load). Should be a good ballpark at least, FWIW.

[Jay_Diddy_B]

Hi,

Here is a reference to efficiency of power distribution transformers:



Taken from:

https://documents.hammondpowersolutions.com/documents/Literature/Medium_Voltage_Distribution_Transformers/HPS-Millennium-G-VPI-Brochure.pdf


This is Hammond Power Solutions.

This is only resistive losses, it is not regulation.

Notice how the efficiency improves as at the larger kVAs.

Regards,
Jay_Diddy_B

[engrguy42]

Jay_Diddy_B,
Not sure why you're focusing on efficiency and power factor...

I assumed the major issue is the harmonic currents causing voltage drops and voltage distortion.

What am I missing?

And the resistive components of the power system equipment are relatively tiny compared to inductive.

[Jay_Diddy_B]

Jay_Diddy_B,
Not sure why you're focusing on efficiency and power factor...

What am I missing?

And the resistive components of the power system equipment are relatively tiny compared to inductive.

You have to consider PF and efficiency in order to estimate the impedance of the source.

Efficiency alone will only give you the resistive component.

Consider this phasor diagram, not drawn to scale:



The regulation is the relatively lengths of the vectors Vin and Vout

The resistive loss in the source is parallel to Iout.

The inductive component in the source is perpendicular to IR

So you need to make an assumption about PF to arrive X_L from the regulation.

Regards,
Jay_Diddy_B

[engrguy42]

Jay_Diddy_B,

Yeah, technically you're correct, no doubt. I just think that in practice you can pretty much ignore the minor resistive components when you're ballparking like this.

No biggie...

[drussell]

Jay_Diddy_B,
Not sure why you're focusing on efficiency and power factor...

I assumed the major issue is the harmonic currents causing voltage drops and voltage distortion.

What am I missing?

Power factor is essentially indirectly describing the magnitude of distortion.  When you un-link the "current being drawn waveform" from the "voltage supplying it waveform", you start to introduce distortion in the way the power is being drawn from the source, the power factor of the load reduces from 1.0 as you move away from purely resitive where the current drawn directly tracks the voltage input.  This does not necessarily change the overall efficiency, but it does change how the power is being drawn from the line.

Essentially, if the power factor of your loads is 1.0, your incoming power sine wave will still look like a sine wave.  If your power factor is 0.1, you're not going to be looking at a sine wave anymore, even if it's still a sine wave coming out of the generator.  You can computer model this like above, or physically simulate it on your bench to see for yourself.

This is very easy to play with safely on your bench using a decent low voltage transformer, (preferably a torroid) off the mains, or a generated 50/60 Hz sine wave signal if you have a AF signal generator (optionally followed by a power amp or buffer stage, etc.) and then add some appropriate small series resistance to simulate at least the main resistive component of the grid impedance, then attach your various loads.  You can then monitor visually on your scope and take samples and do calculations based on what you see, simulating what various loads to to the real mains using that low-voltage setup.

The first time I ever realized this was trying to run an inverter/charger off a portable generator at an off-grid cabin.  The battery charger wouldn't put out nearly as much power as you would think was possible, even less than with the previous, smaller generator.  The solution was for me to modify the generator, removing its ability to do 240v output but doubling the stiffness of the 120v output by simply directly paralleling the output windings.

I also added a fairly beefy motor-run capacitor across the line right at the generator (with a fuse, of course) to help improve the power factor (by helping to boost those flattened peaks of the sinewave caused by the rectifier -> battery load, especially when trying to equalize the cells in the battery bank where you need every volt you can get to try to push them up to 15-15.5v across each series string or whatever.)

After the modification, the charging cables to the battery bank (IIRC, I think it was 00 (2/0) AWG, which is about 67.5  mm2) got noticeably warm.  My friend, who is an electrical engineer, came out of the cabin to the generator shed and said "Wow, that's impressive," that he could feel the cables getting warm.  Sure enough, I went and checked and, yep, they were noticeably warm, even though they were only doing something like 125A.  Impressive, though, indeed, since we could barely get above 30A before, even after doubling the conductor cross-sectional area from the generator shed to the inverter/charger.

I hadn't ever really considered that stuff before because most of the time, for us mere mortals and our typical residential loads, you can just assume that the grid is infinitely low impedance and will always be able to supply your juice, regardless of how "bad" your load is.  This is not actually the case in the macro picture and most people, even many electricians and many, many engineers do not fully grok this, unless they actually work with power electronics or distribution systems on a regular basis.

Edit: Added additional explanation in first paragraph for clarity.

[Simon]

Simon,
Perhaps the attached might help? A discussion of voltage distortion.

It gives a nice perspective of your house load relative to the "grid", and clearly states my point. 

I give up! Just for the record You want to unplug 3 computers and see an appreciable change when the waveform you see in the first place is the result of hundreds of thousands of devices like the mere 3 you unplugged dotted around on the grid between you and the generator. I just give up!

[engrguy42]

Simon,
Perhaps the attached might help? A discussion of voltage distortion.

It gives a nice perspective of your house load relative to the "grid", and clearly states my point. 

I give up! Just for the record You want to unplug 3 computers and see an appreciable change when the waveform you see in the first place is the result of hundreds of thousands of devices like the mere 3 you unplugged dotted around on the grid between you and the generator. I just give up!

Simon,
I would suggest you look at Jay_Diddy_B's excellent simulation above to show the point we're trying to make (that a small rectified load in your home can affect your mains), but it sounds like your mind is made up.

[drussell]

Oh, and I recall a typical distribution transformer impedance is somewhere around 5%

Impedance is not something that is measured or specified as a percentage.

Yeah, technically you're correct

Isn't that pretty much the only kind of correct? 

[engrguy42]

FWIW, I tried to dupiicate Jay_Diddy_B's excellent LTSpice simulation, with some tweaks to make it more appropriate for a US utility service.

The below image shows the circuit and the measured mains voltage. Now of course the assumptions going in can be limitless in terms of the loads and their characteristics and quantities and so on, so anyway...

I assumed at single phase of a typical 12kV distribution circuit that is 2 miles from the station, feeding a 7200/120 pole mounted transformer and the secondary going directly into your house. I applied a 500watt resistive load, and the same full wave rectifier to simulate some AC/DC converter which feeds a filtered, constant current load (in this case 1amp).

And the result is a very gnarly mains waveform. And if I drop the DC load to 0.2 amps like Jay_Diddy_B has the waveform looks a lot like his.

So I suppose a raw, full wave rectified load in your house that draws up near an amp can cause a nasty waveform?

[engrguy42]

Oh, and I recall a typical distribution transformer impedance is somewhere around 5%

Impedance is not something that is measured or specified as a percentage.

Yeah, technically you're correct

Isn't that pretty much the only kind of correct? 

Yes, utility transformer impedance is specified on the nameplate in %. Look it up.

And there is technically correct but irrelevant. If something doesn't matter, it may be true but not worthy of spending time to include it. It's one of those engineering things.

EDIT: Here, I saved you the trouble. Top left corner.

[Simon]

Simon,
Perhaps the attached might help? A discussion of voltage distortion.

It gives a nice perspective of your house load relative to the "grid", and clearly states my point. 

I give up! Just for the record You want to unplug 3 computers and see an appreciable change when the waveform you see in the first place is the result of hundreds of thousands of devices like the mere 3 you unplugged dotted around on the grid between you and the generator. I just give up!

Simon,
I would suggest you look at Jay_Diddy_B's excellent simulation above to show the point we're trying to make (that a small rectified load in your home can affect your mains), but it sounds like your mind is made up.

That is not what he said. Are you next door to a power station? If not there is significant impedance between you and the power station then "things" in the middle will control what arrives at your house. So you think unplugging 3 computers will change the waveform you are already receiving? You know this is not the case. You have already admitted that measuring with and without makes no difference but you have decided it should make a difference. You have your own evidence of reality and you won't believe it!

Yes the simulation shows what I have been saying. Now if you can get everyone in your street to turn everything but resistive loads off you might make a difference.

This is truly hilarious, 100kW of supply coming in and you think changing 300W of the load will change what it looks like. The grid is full of non PFC loads, many of them are contrary to your incorrect assumption not following the rules. The grid will be able to tolerate x% of non power factor corrected devices before it all goes up in flames, that is a judgement made by the rule about the maximum a non PFC load can be which produces the distortion you see which is deemed to be tolerable.

Phone chargers, LED lights, battery chargers, monitors, tv, boiler parts, washing machine parts, laptops, soldering irons, routers, fax machines, inkjet printers, stereos, dishwasher electronics and small parts, toothbrushes, razors, rechargeable sex toys, oscilloscopes, bench multimetors - most stuff in your lab, externally powered USB hubs you name it, how many 50ish watt devices are there around? tons, tons, and even if they are supposed to be PFC they may not be done properly.

If you look at Daves latest video on light panels he opened the supply, oh look, what did it have? in inline inductor for PFC, how god are those? do they just fix it or do they take the worse out bust still leave some distortion.

[drussell]

I would suggest you look at Jay_Diddy_B's excellent simulation above to show the point we're trying to make (that a small rectified load in your home can affect your mains), but it sounds like your mind is made up.

A small rectified load in your home will not appreciably, detectably, affect the actual mains, though.  No.

If you have, for example, many computers at the end of a long extension cord or otherwise "poorly" connected to the mains, it will look worse on a scope at that load end, yes.  Indubitably.  Will those computers on that one extension cord actually appreciably affect the mains grid?  No.  Not even if you measure back at your distribution panel, (say, by monitoring a different, totally unloaded circuit that is on the same leg of the split-phase transformer (assuming 120/240 residential) or same phase on a 3-phase supply.)  Will the sum of all of the computers connected to your local grid make distorted from pure sine at various arbitrary points along the grid?  Absolutely. 

The closer you are to the actual generator you are measuring, the cleaner it will look.  The closer you are to the "messy" load, the "worse" it will undoubtedly look / measure.

[drussell]

Yes, utility transformer impedance is specified on the nameplate in %. Look it up.

I thought that was just used for calculating circulating currents for things like fault-condition calculations, etc, like if you have a 10% 120v transformer, 12v would cause full rated current with a shorted secondary....

...but I'm not a power systems engineer, so I'll let those more knowledgeable in the field chime in here.

[engrguy42]

I would suggest you look at Jay_Diddy_B's excellent simulation above to show the point we're trying to make (that a small rectified load in your home can affect your mains), but it sounds like your mind is made up.

A small rectified load in your home will not appreciably, detectably, affect the actual mains, though.  No.

If you have, for example, many computers at the end of a long extension cord or otherwise "poorly" connected to the mains, it will look worse on a scope at that load end, yes.  Indubitably.  Will those computers on that one extension cord actually appreciably affect the mains grid?  No.  Not even if you measure back at your distribution panel, (say, by monitoring a different, totally unloaded circuit that is on the same leg of the split-phase transformer (assuming 120/240 residential) or same phase on a 3-phase supply.)  Will the sum of all of the computers connected to your local grid make distorted from pure sine at various arbitrary points along the grid?  Absolutely. 

The closer you are to the actual generator you are measuring, the cleaner it will look.  The closer you are to the "messy" load, the "worse" it will undoubtedly look / measure.

Ah, so we agree !! Of course it depends on your definition of "small". The point is (and the one that Simon can't accept) that stuff in your house can affect the distortion of the voltage waveform ("mains") you measure at your wall outlet on your scope. It's the same point outlined in that document I posted, and the same point that is clear when you do as Jay_Diddy_B and I did with an LTSpice simulation.

[Jay_Diddy_B]

Hi,
I can extend my model to show what happens in a Neighbourhood or Neighborhood (perhaps I should say 'street'?):



Assume that there are 10 other house attached to the same distribution transformer. On average they have a 90% / 10% split of resistive and rectifier loads.
In your house you turn-off all the rectifier loads, leaving only the resistive loads, or loads that draw sinewave current.

Results



There is, of course, very little change because your neighbour's loads have distorted the power coming from the source.

Regards,
Jay_Diddy_B

[drussell]

Ah, so we agree !! Of course it depends on your definition of "small". The point is (and the one that Simon can't accept) that stuff in your house can affect the distortion of the voltage waveform ("mains") you measure at your wall outlet on your scope.

He didn't say that.  Absolutely you can affect the readings you get right at your load, the magnitude depending on how heavy and distorted the load is, and how well connected to the grid it is. 

Turning off a couple relatively light, probably actually fairly clean PF-wise loads that are fairly well connected via 14ga/12ga wire to your panel which is then connected via typically 6ga or bigger out to a transformer that is well connected to the grid through a relatively low impedance source, very stiff, with massive fault current capability, won't show much difference, even right at your load.

You were talking about affecting the mains coming into your premise.  You're not likely going to be able to do that appreciably with a few small computer loads, even if they don't have active PFC. 

Now, on the other hand, if you have a 30 HP pressure washer that you could fire up.....